Brief Details: NSQL-Llama-2-7B is a specialized SQL generation model based on Llama-2, fine-tuned on 1M SQL queries and text-to-SQL pairs for precise query generation.
BRIEF-DETAILS: A Stable Diffusion model fine-tuned on Genshin Impact landscapes using DreamBooth. Creates anime-style landscapes with "ggenshin landscape" prompt.
Brief Details: Lumina-T2I is a powerful text-to-image generation model using LargeDiT backbone with LLaMA-7B text encoder and SDXL VAE, supporting 1024x1024 resolution.
Brief-details: Alpaca-30B is a LoRA-fine-tuned LLaMA model trained on the Tatsu Labs Alpaca dataset, optimized for instruction-following tasks with 8-bit quantization
Brief-details: K2 is a 65B parameter LLM that outperforms Llama 2 70B using 35% less compute, trained on 1.4T tokens with Apache 2.0 license and robust evaluation metrics.
Brief Details: A 46.7B parameter LLM fine-tuned from Mixtral-8x7B, specialized in reasoning tasks with native chain-of-thought capabilities and competitive performance.
Brief-details: LLaMA2-13B-Tiefighter is a creative merged model combining multiple LORAs for enhanced story writing, chatbot interactions, and adventure gaming capabilities.
Brief-details: Command-R 35B GGUF - High-performance 35B parameter model optimized for llama.cpp, featuring F16 quantization and split file support
Brief-details: Cross-domain Chinese embedding model (400MB) with SOTA performance on MTEB benchmark. Supports 1024 token context window, optimized for search, Q&A, and RAG applications.
Brief-details: Accelerated text-to-image model using LCM-LoRA technology on SSD-1B base, enabling fast 2-8 step inference with 105M parameters and OpenRAIL++ license.
Brief Details: ShareGPT4V-7B is a multimodal chatbot combining CLP vision and LLaMA/Vicuna, trained on 1.2M image-text pairs for enhanced visual-language understanding.
Brief-details: A powerful 34B parameter code generation model quantized to GGUF format, achieving 73.2% pass@1 on HumanEval, surpassing GPT-4's early 2023 performance.
Brief Details: A specialized image tagging model based on MOAT architecture, trained on Danbooru dataset with strong F1 score of 0.6911. Supports ratings, characters, and general tags.
Brief-details: A 15.5B parameter code generation model fine-tuned on GPTeacher dataset, combining StarCoder's capabilities with instruction-following abilities for programming tasks.
Brief-details: A specialized text-to-image diffusion model focused on high-quality backgrounds and detailed expressions, featuring built-in VAE and multiple versions optimized for different use cases.
BRIEF DETAILS: ChatGLM-6B-INT4-QE is a quantized bilingual LLM with 6B parameters, optimized for Chinese-English dialogue. Features 4-bit quantization for efficient 6GB deployment.
Brief Details: Hybrid 7.5B parameter model combining Mamba2 state-space and transformer blocks, optimized for instruction-following with 16k context window
Brief-details: A large-scale Chinese MRC (Machine Reading Comprehension) model based on RoBERTa-wwm-ext architecture, achieving state-of-the-art performance on Dureader-2021 and medical QA tasks.
Brief-details: A creative AI model mix utilizing CLIP 2 and WD 1.4 VAE for image generation, developed by DrBob2142 with 80 community likes, focusing on diverse visual outputs.
Brief Details: A specialized ControlNet SDXL model for converting simple sketches into high-quality anime illustrations, featuring improved aesthetic scores and robust prompt following capabilities.
Brief-details: A 2.51B parameter multilingual LLM optimized for 10 Indic languages + English, trained on 2T tokens with superior tokenization efficiency for Indic scripts
